Around 60 people turned up to the Apple Store’s auditorium last night to attend our joint event with the Royal Institution on the science of search. (Although you never know how many were actually there for the free wifi.)

Three speakers covered very different aspects of search. Ewan Birney, a senior scientist at EMBL near Cambridge, talked about methods to search and compare genomes. How do you take several billion base pairs from one species and compare them with similarly monstrous amounts of data from other species? Not without melting the power cables to your supercomputer, it turns out. Ewan was keen to emphasise the coming problems with public perception of genetics. Over the next few years, science will uncover genetic risk factors for ever more diseases. But the increased risk from a particular genotype is often tiny. Will individuals who learn of these slight predispositions understand the odds, given the tendency of the press to exaggerate or sensationalise biomedical results?

Nature’s Timo Hannay gave an overview of some of the methods used to mine literature. How can a computer program extract data from the vast scientific literature when publishers all use different layouts for that information and often put it behind a subscription wall. The Open Text Mining Interface (OTMI) is one solution (and also happens to be an anagram of the speaker’s name). Timo concluded with a look at Connotea, Nature’s social bookmarking site for scientists.

Jonathon Hare from Southampton University dusted things off with a fascinating tour of image searching. We’re all familiar with Google Images, and the wildly irrelevant results it often throws up. Jonathan presented alternatives to search based on text surrounding images. For example, search by sketch has the user drawing a rough sketch of the object they’re after – a rose, say. An algorithm then searches based on shape and colour. The method has obvious limitations. How, for example, would one meaningfully draw ‘agility’ or ‘George Bush shaking hands with Tony Blair’? But other tricks are being developed that should radically improve image search in the next five years.

Thanks to all the speakers for enlightening talks, and to Bill Thompson – a Chair with flair and lots of hair.